Recurrent laryngeal nerve (RLN) injury after thyroidectomy is relatively common. Locating the RLN prior to thyroid dissection is paramount to avoid injury. We developed a fluorescence imaging system that permits nerve autofluorescence. We aimed to determine the sensitivity and specificity of fluorescence imaging at detecting the RLN relative to thyroid and other background tissue and compared it to white light.

In this prospective study, 65 patients underwent thyroidectomy from January to April 2022 (16 bilateral thyroid resections) using white and fluorescent light. Fluorescence intensity [relative fluorescence units (RFU)] was recorded for RLN, thyroid, and background. RFU mean, minimum, and maximum values were calculated using Image J software. Thirty randomly selected pairs of white and fluorescent light images were independently reviewed by two examiners to compare RLN detection rate, number of branches, and length and minimum width of nerves visualized. Parametric and nonparametric statistical analysis was performed.

All 81 RNLs observed were visualized more clearly under fluorescence (mean intensity, µ = 134.3 RFU) than either thyroid (µ = 33.7, p < 0.001) or background (µ = 14.4, p < 0.001). Forest plots revealed no overlap between RLN intensity and that of either other tissue. Sensitivity and specificity for RLN were 100%. All 30 RLNs and all 45 nerve branches were clearly visualized under fluorescence, versus 17 and 22, respectively, with white light (both p < 0.001). Visible nerve length was 2.5 × as great with fluorescence as with white light (µ = 1.90 vs. 0.76 cm, p < 0.001).

In 65 patients and 81 nerves, RLN detection was markedly and consistently enhanced with autofluorescence neuro-imaging during thyroidectomy, with 100% sensitivity and specificity.

Autofluorescence neurography; Fluorescence imaging; Nerve autofluorescence; Prospective study; Prototype imaging system; Recurrent laryngeal nerve; Thyroidectomy.